|
NatNumDistribWithTail.java
|
/* * Copyright (c) 2005, 2006, Regents of the University of California * All rights reserved. * * Redistribution and use in source and binary forms, with or without * modification, are permitted provided that the following conditions * are met: * * * Redistributions of source code must retain the above copyright * notice, this list of conditions and the following disclaimer. * * * Redistributions in binary form must reproduce the above copyright * notice, this list of conditions and the following disclaimer in * the documentation and/or other materials provided with the * distribution. * * * Neither the name of the University of California, Berkeley nor * the names of its contributors may be used to endorse or promote * products derived from this software without specific prior * written permission. * * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS * "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT * LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS * FOR A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE * COPYRIGHT OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, * INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES * (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR * SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, * STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED * OF THE POSSIBILITY OF SUCH DAMAGE. */ package blog.distrib; import java.util.*; import Jama.Matrix; import blog.*;
| A mixture of an explicit distribution over the first k natural numbers 0,...,(k-1), and a geometric distribution over the numbers greater than or equal to k. This distribution takes three parameters: a column vector of k probabilities for the first k numbers, the probability lambda that the value is less than k, and the success parameter alpha of the geometric distribution. |
public class NatNumDistribWithTail extends AbstractCondProbDistrib {
Creates a NatNumDistribWithTail with the given array of probabilities
for the multinomial distribution, and the given lambda and alpha
values. k is set to the size of the probability array.
distribution over the first pi.length natural numbers
P(X >= n+1 | X >= n) for n >= pi.length
|
public NatNumDistribWithTail(double[] pi, double lambda, double alpha) {
k = pi.length;
prefixDistrib = new Categorical(pi);
mixDistrib = new Bernoulli(lambda);
geometric = new Geometric(alpha);
}
Creates a new NatNumDistribWithTail with the following three
parameters:
|
public NatNumDistribWithTail(List params) {
if (params.size() != 3) {
throw new IllegalArgumentException
("NatNumDistribWithTail expects three parameters: the "
+ "distribution over small numbers, the probability of "
+ "using that distribution, and the success probability "
+ "for a geometric distribution over larger numbers.");
}
if ((!(params.get(0) instanceof Matrix))
|| (((Matrix) params.get(0)).getColumnDimension() != 1)) {
throw new IllegalArgumentException
("First parameter to NatNumDistribWithTail should be a "
+ "column vector.");
}
Matrix m = (Matrix) params.get(0);
k = m.getRowDimension();
double[] probs = m.getColumnPackedCopy();
prefixDistrib = new Categorical(probs);
if (!(params.get(1) instanceof Number)) {
throw new IllegalArgumentException
("Second parameter to NatNumDistribWithTail should be a "
+ "number (the probability of using the explicit distrib).");
}
mixDistrib = new Bernoulli(((Number) params.get(1)).doubleValue());
if (!(params.get(2) instanceof Number)) {
throw new IllegalArgumentException
("Third parameter to NatNumDistribWithTail should be a "
+ "number (the success prob for the geometric distrib).");
}
geometric = new Geometric(((Number) params.get(2)).doubleValue());
}
| Returns the probability of a non-negative integer n under this distribution. Throws an exception if n is negative. |
public double getProb (int n) {
if (n < k) {
return mixDistrib.getProb(true) * prefixDistrib.getProb(n);
}
return mixDistrib.getProb(false) * geometric.getProb(n - k);
}
| Returns the log probability of a non-negative integer n under this distribution. Throws an exception if n is negative. |
public double getLogProb (int n) {
if (n < k) {
return mixDistrib.getLogProb(true) + prefixDistrib.getLogProb(n);
}
return mixDistrib.getLogProb(false) + geometric.getLogProb(n - k);
}
| Returns the probability of the given value, which should be a non-negative Integer. Expects no arguments. |
public double getProb(List args, Object childValue) {
if (!args.isEmpty()) {
throw new IllegalArgumentException
("NatNumDistribWithTail expects no arguments.");
}
if (!(childValue instanceof Integer)) {
throw new IllegalArgumentException
("NatNumDistribWithTail defines distribution over objects "
+ "of class Integer, not " + childValue.getClass());
}
return getProb(((Integer) childValue).intValue());
}
| Records an occurrence of the number n, for use in updating parameters. |
/*
public void collectStats(int n) {
if (n < 0) {
throw new IllegalArgumentException
("MultinomialWithTail can't generate a negative number.");
}
totalCount++;
if (n < k) {
multinomialCount++;
multinomial.collectStats(n);
} else {
geometric.collectStats(n-k);
}
}
*/
| Returns a sample from this distribution. |
public int sampleVal()
{
boolean usePrefix = mixDistrib.sampleVal();
if (usePrefix) {
// X < k
return prefixDistrib.sampleVal();
}
else {
// X >= k
return k + geometric.sampleVal();
}
}
| Returns an Integer sampled from this distribution. Expects no arguments. |
public Object sampleVal(List args, Type childType) {
if (!args.isEmpty()) {
throw new IllegalArgumentException
("NatNumDistribWithTail expects no arguments.");
}
return new Integer(sampleVal());
}
| Sets the parameter lambda to the value that maximizes the likelihood of the numbers passed to collectStats since the last call to updateParams. Then clears the collected statistics, and returns the difference between the log likelihood of the data under the new parameters and the log likelihood under the old parameters. |
/*
public double updateParams() {
// Update mixture parameter
double oldLogProb = (multinomialCount * logLambda) +
((totalCount - multinomialCount) * logOneMinusLambda);
if (totalCount > 0) {
lambda = multinomialCount / (double) totalCount;
cacheParams();
}
double newLogProb = (multinomialCount * logLambda) +
((totalCount - multinomialCount) * logOneMinusLambda);
// Update parameters for component distributions
double logProbChange = newLogProb - oldLogProb;
logProbChange += multinomial.updateParams();
logProbChange += geometric.updateParams();
// Clear statistics
totalCount = 0;
multinomialCount = 0;
return logProbChange;
}
*/
int k;
Bernoulli mixDistrib;
Categorical prefixDistrib;
Geometric geometric;
}
This file was generated on Tue Jun 08 17:53:47 PDT 2010 from file NatNumDistribWithTail.java
by the ilog.language.tools.Hilite Java tool written by Hassan Aït-Kaci